When the General Model Is Not Enough
The ten guides beside this one cover generalists. This one covers the specialists: purpose-built tools that beat a horizontal assistant inside one domain, in coding, legal, finance, customer service, enterprise knowledge and healthcare. The recurring question is not “which is smarter,” it is “does this job clear the bar where a specialist’s data, workflow and guardrails are worth the price, the lock-in and the integration cost?”
A different question from “which model is best”
General assistants are extraordinary generalists. But in regulated, data-heavy or workflow-bound work, a tool that embeds the domain, the case law, the codebase, the EHR, the data room, often beats a smarter but generic model. The 2026 market has a credible specialist for most professional functions. The skill is knowing when to reach for one.
Domain layer over a frontier model
Most vertical tools run on the same OpenAI or Anthropic models as your general assistant. What you pay for is the layer on top: proprietary data, workflow, integrations, evaluation and accountability.
- The intelligence is often rented
- The domain layer is the product
- Judge the layer, not the demo
Depth a generalist cannot fake
A specialist embeds the things a general chat cannot: your repository, your matter history, your permission model, your compliance posture and the workflow the work actually follows.
- Context that lives in your systems
- Guardrails for regulated work
- Integrations into systems of record
Price, lock-in and overlap
Specialists cost more, embed your data deeply, and often overlap with the general assistant you already pay for. The decision is economic and architectural, not just about capability.
- Your Copilot may already do 80%
- Migration is rarely easy
- Buy only for the part that clears the bar
A specialist for most professional functions
This is the shape of the market in late 2026: a well-funded, credible tool for most domains a general assistant handles only shallowly. Names change fast, and so do the boundaries, so treat this as a snapshot of categories rather than a permanent leaderboard.
| Domain | What the specialist adds | Leading examples |
|---|---|---|
| Coding | Repository context, autonomous multi-file agents, CI and pull-request integration | Cursor (now owned by SpaceX), Cognition (Devin) |
| Legal | Case law, contracts, firm knowledge, citation discipline and conflict-of-interest enforcement | Harvey |
| Finance and research | Deep analysis across private documents, public filings and licensed market data, now delivered into other tools over MCP | Hebbia |
| Customer service and long-horizon work | Outcome-priced agents across chat, voice, email and messaging, now extending to goals that run for weeks | Sierra |
| Enterprise knowledge | Permission-aware search and agents across all company apps, with agents that carry their own identity | Glean |
| Healthcare | Ambient clinical notes inside the EHR, extending into workflow and revenue cycle; evidence at the point of care | Abridge, OpenEvidence |
The most mature vertical, and the most contested
Software is where vertical AI is furthest along and where the general assistants also compete hardest. Claude Code and OpenAI Codex are the generalist coding agents; Cursor and Cognition are the purpose-built ones. The line between “feature of my assistant” and “dedicated tool” is blurriest here, and this window shows the two vendors moving at very different speeds: Cursor shipped three times, Cognition published nothing at all.
Cursor IDE agent · SpaceX-owned
Positions itself as a coding agent for building ambitious software: deep codebase understanding, autonomous and parallel agents, and bring-your-own-model across OpenAI, Anthropic, Gemini and others. Its enterprise page claims 64% of the Fortune 500 and more than 50,000 enterprises, with SOC 2 Type II, SAML and SCIM (vendor claims). The ownership change is confirmed on Cursor’s own blog, dated August 14, 2026 and titled “Cursor is now a part of SpaceX”: “Cursor has officially been acquired by SpaceX.” No valuation and no terms are disclosed, so do not put a number on it. Cursor Router, released July 22, adds model routing in Intelligence, Balance and Cost modes, which Cursor says delivers frontier-quality performance at 60% savings (vendor claim), though the docs are explicit that “Cursor Router is currently only available on Teams and Enterprise plans,” so it is a policy lever for an organization rather than something an individual developer can switch on. Shipping continued through the window: cloud agents that run on machines you manage (September 2), starting a project from scratch without a repo (August 27) and Cloud Agents plus Cursor Harness improvements (August 19), after an August run that included Git at any scale, Origin Code Hosting in early beta and AIUC-1 certification.
- Acquired by SpaceX on August 14, 2026, no terms disclosed
- Repo-aware, multi-model, still shipping weekly
- Router is Teams and Enterprise only
- Ownership is now a procurement question, not a footnote
Cognition (Devin) Autonomous SWE
Operates Devin, marketed as the first autonomous software engineer: it plans, writes, tests and ships code inside your codebase and tools. Cognition has moved toward named capabilities rather than version numbers, with Devin 2.2 the last numbered release still cited below: Devin Fusion, described on June 29, 2026 as a new kind of multi-model harness, and Devin Security Swarm. On FedRAMP, read the exact words, unchanged on September 3, 2026: “Cognition’s entire platform is now FedRAMP Class D (High) In-Process and listed on the FedRAMP Marketplace,” covering Devin Desktop (formerly Windsurf), Devin Cloud and the Devin CLI. In-Process is a place in a queue, not an authorization, and a public-sector buyer should treat it that way. A July 22 memorandum of understanding with the US Department of Energy joins the Genesis Mission across 17 national labs, alongside existing work with the US Army, Navy and NASA JPL. Cognition is also doing the buying: TierZero closed, and The Interaction Company of California reads as absorbed rather than pending. The notable fact this cycle is silence: Cognition has published nothing since July 28, 2026, the LTM partnership post, which is distribution rather than a product change. That is now a five-week quiet stretch beside a competitor shipping weekly.
- Delegated, end-to-end tasks
- FedRAMP Class D (High) In-Process: a queue position, not an authorization
- Nothing published since July 28, 2026
- Review every diff it ships
The generalist option Already yours
Claude Code and OpenAI Codex bring capable agentic coding inside tools you may already pay for. For many teams this covers the majority of the work before a dedicated tool is justified.
- No new vendor
- Strong for most tasks
- Benchmark against the specialists
Where domain data is the moat
In legal, finance and enterprise knowledge, the value is not raw intelligence. It is grounded access to the right documents, under the right permissions, with the right citation and audit discipline. These tools compete on the corpus and the controls, not the model. Late August 2026 was the busiest fortnight this section has recorded: Glean ran a six-post launch wave, Hebbia shipped a flagship upgrade that distributes itself into other vendors’ tools, and Harvey shipped memory, horizon scanning and a new frontier model on the day it launched.
Harvey Legal
Agents built for law firms and enterprise legal teams, grounded in case law, contracts and firm knowledge. Reports more than 100,000 lawyers across 1,300 organizations, and raised at an $11B valuation in March 2026, a figure that still dates from March 25, 2026 with nothing since restating it. The July 28, 2026 strategic investment from Goldman Sachs and J.P. Morgan publishes no new valuation, so do not infer one. Harvey is the most active vendor on this map: seventeen posts in the window to September 3, 2026. The substantive ones are Fable 5.1 going live in Harvey (September 1, the same day Anthropic launched it), Horizon Scanning (September 1), a write-up of rebuilding Playbook Review as a multi-agent system (September 2), “Memory is Here: Harvey, Personalized” (August 26) and the Harvey Tenet research preview (August 20). Harvey II, announced August 18, still has no general-availability date and no pricing: the post says only “Today, we’re introducing Harvey II” and that “Harvey’s agents now inherit the context of your matters and projects, along with a memory of how you work.” The call to action is “Request a Demo.” Availability is unconfirmed and it does not belong in a dated rollout plan.
- Fable 5.1 live in Harvey on September 1, Anthropic’s launch day
- Harvey II still has no GA date and no pricing
- The $11B valuation still dates from March 25, 2026
- A lawyer still signs the work
Hebbia Finance
Describes itself as the leading AI platform for finance. Its Matrix interface, with Skills and Agents, runs analysis across private documents plus public filings and licensed third-party market data: connectors include SEC filings, CapIQ, FactSet, PitchBook, Preqin, expert networks, Snowflake and SharePoint. The big move is August 26, 2026: Matrix 2.0, described by Hebbia as “a major upgrade to Matrix, Hebbia’s flagship product.” Two claims in it matter. First, output: “Matrix now creates deliverables, including memos, decks, models, and emails.” Second, distribution: “Matrix is now available directly in Claude, ChatGPT, Cursor, or any internal application through the Hebbia MCP.” A finance specialist that arrives inside the generalist you already run is a different procurement conversation from one you log into. Also in the window: Daloopa is now available in Hebbia (September 1) and a “What’s New: September Disclosure 2026” post (August 31). Hebbia Max remains not generally available: the September disclosure calls Max one of “two major launches” of the prior month but never states general availability and never retracts the July 30 line that “Max is rolling out to a small set of firms first.”
- Matrix 2.0 shipped August 26, 2026
- Distributed into Claude, ChatGPT and Cursor over MCP
- Max GA still unconfirmed, still a gated preview
- Verify every extracted figure
Glean Enterprise knowledge
Work AI that unifies permission-aware search, an assistant and agents across 100-plus company apps, so answers respect who is allowed to see what. August 26, 2026 was a coordinated launch wave, six posts the same day: an “intelligence era” platform post on efficient, governed and proactive AI powered by enterprise context; a proactive AI suite for single and multiplayer work; Glean Agents working independently, building faster and staying governed at scale; Glean Transform, positioned as “your company’s mission control for AI transformation”; AI usage controls for scaling usage without runaway spend; and one competitive post. The model picker has moved past this guide’s previous baseline: Gemini 3.7 Flash was added on August 13 alongside GPT-5.6 Luna powering Fast Mode, GPT-5.6 Luna, Sol and Terra were added on August 12, and Claude Opus 5 dates to July 29. Admins also gained control over agent default models on August 13. Glean’s release notes have published nothing since August 18, 2026.
- Search across all your tools, permission-aware by design
- Six-post launch wave on August 26, 2026
- Model picker refreshed August 12 and 13
- Only as good as your data hygiene
Agents got an identity of their own, and it is now GA
Agent Identity entered public beta on July 15, 2026 and reached general availability on August 12, 2026. The previous edition of this guide marked its status unconfirmed; it can now be stated plainly. An agent can “use its own scoped service credentials instead of borrowing the identity of the person who invoked it,” with admin controls for credential scoping and audience management. It remains the most interesting governance idea in this set: access rights split from invocation rights, and audit trails that name the agent rather than whoever triggered it.
- GA on August 12, 2026, no longer unconfirmed
- Scoped service credentials per agent
- Admin controls for scoping and audience
- Audit trails name the agent
A vendor benchmark against a named competitor
One of the six August 26 posts is titled “Right-sizing intelligence: Glean saves 81% on token costs and is preferred 78% of the time versus Claude Cowork.” Both figures are Glean’s own claims about a named competitor, produced by Glean, published by Glean, in Glean’s comparative marketing. This guide records that the post exists and what it asserts; it does not endorse either number. If a token-cost or preference figure is going to influence a purchase, run the comparison on your own corpus and your own queries, because a vendor’s framing of “right-sizing” is a claim about which jobs need which model, and that judgment is exactly what you are buying.
- 81% token-cost saving: Glean’s claim
- 78% preference rate: Glean’s claim
- Vendor comparative marketing, not an independent test
- Reproduce it on your own workload before citing it
A specialist arriving inside a suite
Harvey appeared as a new connector in Microsoft 365 Copilot in Microsoft’s August 2026 roundup. This is the dynamic this guide exists to track: the specialist and the generalist are no longer alternatives on a shortlist, they are layers, and the specialist can reach the user through the suite the organization already governs. Ask who owns the connector, whose permission model applies inside it, and where the audit trail lands. Ethical Wall enforcement with Intapp, generally available since July 23, is the reason that question matters for a law firm: conflict policy has to travel with the tool into whatever surface it appears on.
- Harvey is now a Microsoft 365 Copilot connector
- Two permission models meet in one workflow
- Confirm walls and audit coverage in the connected surface
Beyond your own documents
The July 10 integration post remains the clearest statement of scope: the corpus is no longer just the data room. Private files sit alongside public filings and licensed market data in a single view, which changes the diligence question from “can it read our documents” to “which third-party licenses do we already hold, and which does this tool require of us.” Matrix 2.0 adds a second version of the same question, because the output now leaves through MCP into Claude, ChatGPT and Cursor.
- One view over three kinds of data
- License terms become part of the buy
- MCP distribution adds an egress question
- Source every number back to a filing
For this vertical tool, tell me: which underlying model it uses, exactly what proprietary data or workflow it adds on top, how it handles citations and source traceability, and where our data is stored and processed. Separate vendor marketing from verifiable fact.
Compare what this specialist does against what our existing Microsoft 365 Copilot and ChatGPT already cover. Identify the specific 20% of the workflow the specialist genuinely adds, and whether that 20% justifies the cost and integration.
This vendor published a comparison claiming a cost saving and a preference rate against a named competitor. Design a test I can run on our own documents and queries that would confirm or refute both numbers, and tell me what the vendor’s methodology would have to look like for the claim to hold.
Specialists that act, and specialists that must be right
Two high-stakes frontiers: customer-facing agents that take real actions for real users, and clinical tools where an error has a different weight entirely. Both show why domain guardrails and human accountability matter more than raw capability.
Sierra Agents that act
Began in customer experience, in chat, voice, email and messaging, with outcome-based pricing and a reported 40% of the Fortune 50 as customers, and has now pushed past it. Horizon, announced July 16, targets goals that run for days, weeks or months, such as loan origination and prior authorization, rather than single conversations. A July 14 partnership makes SoftBank Corp. the exclusive sales partner in Japan, where the LINEMO deployment reports 97% resolution and 93% CSAT (vendor figures). Sierra acquired Takeoff on July 23. Founded by Bret Taylor; valued around $15.8B in 2026. The window to September 3, 2026 adds release governance, described as guardrails for agents at scale (August 20), a Korea launch with a Seoul office under the heading “Sophisticated agents for the most demanding companies” (August 25), and a voice AI explainer (September 1). One item is corporate rather than product and should be read that way: Julia Brau Donnelly joined as Chief Financial Officer on August 31, a signal about the company’s stage, not a change to what the product does. No baseline item was retracted.
- Release governance shipped August 20, 2026
- Korea launch and Seoul office, August 25
- CFO hire August 31: a corporate signal, not a product change
- A months-long goal needs a named owner
Abridge Clinical documentation
Turns doctor and patient conversations into structured clinical notes in real time, integrated into Epic. In June 2026 NVIDIA collaborated with Abridge on a clinical-conversation foundation model. The scope is widening past documentation toward care delivery, payment and revenue cycle: the July 20 Altrina acquihire adds browser-based AI agents and EHR workflow automation, framed as a talent investment, and CHRISTUS Health expanded its deployment on July 21. Ambient documentation is still an accurate description, but increasingly names only a subset of the product. The window to September 3, 2026 is customer content rather than product news: a Sutter enterprise-AI timeline (August 27), “Now in Practice: August 2026” and a West Virginia University Medicine piece (both August 28). Nothing was published between September 1 and 3.
- Ambient notes are the beachhead, not the ceiling
- August posts are customer content, not product change
- Clinician reviews every note
- Consent and privacy are non-negotiable
OpenEvidence Clinical evidence
A medical answer engine grounded in licensed clinical literature, with multi-year content agreements with the NEJM Group and JAMA Network and enterprise deployments across major health systems.
- Evidence at the point of care
- Licensed, citable sources
- Clinical judgment stays human
A decision framework, not a vibe
The specialist-versus-generalist and buy-versus-build calls are recurring and expensive. Run them deliberately, and re-run them, because the thing you evaluated in July may arrive through a different surface in September.
A $15B valuation is not a fit for your workflow
Specialist AI carries risks a general assistant does not: you are often paying a premium for someone else’s model, embedding sensitive data deeply, and betting on a market that is consolidating in real time. Six deals have landed across the eight vendors in this guide since July 10, 2026, the latest being Cursor’s acquisition by SpaceX on August 14. The pace is uneven, not a metronome: five deals in seventeen days, then eighteen days with none, then the largest change of control in the set, then nothing again through September 3. Do not annualize a burst, and do not read a quiet stretch as the end of one.
Rented intelligence
Many verticals wrap OpenAI or Anthropic. Confirm the domain layer is real before paying specialist prices for someone else’s model.
Overlap you already own
Your Copilot, ChatGPT or Claude may already cover most of the job. Buy the specialist only for the measured gap. This is getting harder to assess, not easier: Hebbia now runs inside Claude, ChatGPT and Cursor over MCP, and Harvey is a Microsoft 365 Copilot connector, so the same capability can reach a user through two contracts at once.
Vendor comparisons name competitors now
Glean’s August 26 post claims it “saves 81% on token costs and is preferred 78% of the time versus Claude Cowork.” Those are Glean’s figures about a named rival, published in Glean’s own marketing. Treat any head-to-head number from a vendor as a hypothesis to test on your corpus, and be equally skeptical when a competitor answers with numbers of its own.
Data and compliance
Vertical tools touch your most sensitive data: legal, clinical, financial. Verify storage, processing, and HIPAA, SOC 2 or EU AI Act posture. Add one question for MCP distribution: when the specialist runs inside another vendor’s assistant, whose data-processing terms apply?
Lock-in
Proprietary data and workflow make these tools hard to leave. Check export and migration before you embed one in a critical path.
Announced is not available
Two items on this map have been announced without a date for weeks. Harvey II was introduced on August 18, 2026 with no GA date and no pricing, and the call to action is still a demo request. Hebbia Max has been “rolling out to a small set of firms first” since July 30, and the August 31 disclosure calls it a major launch without ever saying it is generally available. Neither belongs in a dated rollout plan.
Model support goes stale in days
Fable 5.1 reached Microsoft 365 Copilot and Harvey on September 1, its launch day, while Glean’s release notes had not recorded it as of August 18. Any capability matrix you circulate should carry the date it was compiled, because a fortnight is long enough to make it wrong.
Consolidation, at speed
Six deals across these eight vendors since July 10, 2026. In the seventeen days to July 27: Cognition bought TierZero and welcomed The Interaction Company, Harvey bought Benchmark, Sierra bought Takeoff, Abridge took on the Altrina team. Then eighteen quiet days. Then, on August 14, Cursor announced that it “has officially been acquired by SpaceX,” completing a process begun in April; no valuation or terms were disclosed. Harvey separately took a strategic investment from Goldman Sachs and J.P. Morgan on July 28, with no new valuation published. The tracked set now contains a vendor owned by a launch company, which sharpens the point rather than softening it: the specialist you sign with may not merely be a different company in a year, it may sit inside a group whose priorities have nothing to do with your codebase.
Do not infer ownership, but do not ignore the signal
This guide’s July 27 edition removed a draft claim that Cursor had been acquired: no source supported it and the date it gave was wrong. On August 14, 2026 Cursor announced the acquisition itself, first-party and dated, in a post titled “Cursor is now a part of SpaceX.” Both calls were right, and the arithmetic is the argument for the rule: the guide was wrong for zero days, and would have been wrong for eighteen. Keep the discipline and note its real price, which is low. A partnership, a joint model launch or a shared brand is evidence of a relationship, not of ownership. Get ownership, control and model supply written into the contract rather than inferred from a launch post, and put a change-of-control clause in anyway, because the inference does sometimes come true.
A missing price is not a discontinued product
Cursor’s pricing page renders four prices and omits Pro+, Ultra and Teams Premium, and it does not list Cursor Start or Cursor for iPad at all. Whether Start was withdrawn or is geo-gated to visitors in India cannot be determined from the page, so this guide marks it unconfirmed. Resist both readings until the vendor says which it is, and get any tier you intend to buy confirmed in a written quote.
Valuation is not fit
A famous, well-funded tool can still be wrong for your one workflow. Evaluate on your job, not on the funding round.
| Before you buy a vertical tool | What to check | Who owns it |
|---|---|---|
| It wraps a frontier model | Which model, and how fast the vendor adopts the next one | Technical evaluation |
| It touches sensitive data | Where data is stored and processed; compliance certifications | Security and legal |
| It overlaps tools you own | The exact gap versus your existing general assistant | Budget owner |
| It arrives through another vendor | Whose permission model, terms and audit trail apply in that surface | Security and procurement |
| It embeds deeply | Data export and migration path if you leave | Procurement |
| It produces work of record | Human review of every filing, diff, note or customer action | The professional of record |
| It may be acquired | Change-of-control terms, and what happens to price, roadmap and support | Procurement and the budget owner |
First-party evidence behind this guide
This is a fast-moving, fast-consolidating category, so these vendor pages are the most volatile sources in the whole set: re-verify before you act on any specific tool. Capability claims are anchored to each vendor’s own pages; valuations, funding and comparative benchmarks are as reported by the vendor and are labeled as such in the text. Read one limit before you cite anything here: this guide lists roughly one representative dated article per vendor, and many other items referenced in the body were read on the vendors’ own channels without being separately cited below. Those include Cursor’s September 2 post on cloud agents running on machines you manage, the August 27 start-from-scratch post and the August 19 Cloud Agents and Harness improvements; Harvey’s Fable 5.1, Horizon Scanning, Playbook Review and Memory posts and the Tenet research preview; Harvey appearing as a Microsoft 365 Copilot connector in Microsoft’s August roundup; Hebbia’s September 1 Daloopa post and its August 31 September Disclosure; Sierra’s release-governance post on August 20, its Korea launch on August 25 and its CFO announcement on August 31; and Abridge’s August 27 and August 28 customer posts. Glean’s model-picker and Agent Identity dates come from Glean’s rolling release notes, which cannot anchor a dated claim, so they are stated in the body with that limitation named. One stated exception is larger than the rest: OpenEvidence is tracked in this guide but carries no first-party citation below at all, so the claims in its card, the multi-year content agreements with the NEJM Group and JAMA Network and the enterprise deployments across health systems, were not re-verified this cycle and a reader relying on them should confirm them with OpenEvidence directly. If you are relying on any one of these specifically, confirm it directly with the vendor.
AI Mindset